Undesired alteration of digital images is complicated due to the relative ease with which the digital image may be reproduced and distributed. Secure communication methods, such as various encryption methods, provide one form of protection for digital images. Briefly, and in general terms, encryption may prevent outsiders from viewing the digital image or video sequence, but once the image or video stream has been decrypted, all protection against manipulation and redistribution of the image or video stream are eliminated. Moreover, if it is desired to distribute the digital image or video sequence on a widespread basis, management of the encryption key in a public key encryption scheme may become cumbersome. In addition, since encryption generally requires error-free communications, distribution of the digital image or video stream may not be available in all networks, particularly wireless networks.
An alternative method for preventing undesired alteration of digital images may be provided by a message authentication code that may be attached to the digital image. For example, in one known method, origination information may appear within a message field appended to the digital image. One significant shortcoming present in this type of add-on authentication is that it is easily identified and removed from the digital image. A more resilient system may be provided by selectively encoding data into the digital image that is generally not perceptible to a viewer.
Digital marking is one known method for providing an identifier in a digital image. Typically, a digital mark is unobtrusively embedded in the image so that an attempted removal of the digital mark generally destroys the underlying digital image. The mark may further be embedded into the original data so that it is generally imperceptible to a viewer. Accordingly, when suitably marked digital signal content is distributed to others, the data included within the mark is distributed with the digital signal content, so that the originator can non-ambiguously demonstrate the source of the content. Digital marking may also permit tampering to be identified, since any manipulation of the data within the mark will result in an invalid mark.
A variety of conventional digital marking methods have been proposed. Typically, these digital marking methods insert a digital mark into the high-frequency content of an image or other product. Although marginally effective with respect to some types of attacks, the conventional digital marking methods are vulnerable if techniques such as low-pass filtering are used. In addition, the conventional digital marking methods tend to be labor-intensive, relatively slow, and lack the ability to be adequately automated.